【作者】 黄艳华；

【导师】 吴贤国；

【作者基本信息】 华中科技大学 ， 工程管理， 2017， 硕士

【摘要】 随着城市轨道交通大规模的建设和轨道交通线网的扩大,地铁逐渐成为目前城市生活中最为方便快捷的出行方式之一。但在内部因素和外部环境的共同作用下,地铁运营逐渐出现各项问题,成为人民和众多学者关注的焦点。我国地铁大规模的建设与运营起步较晚,针对地铁的各项研究体系缺乏完善,将各种智能化数据挖掘方法运用于运营地铁结构健康风险感知与诊断领域中的能力较为薄弱,因此,定性探索运营地铁结构损伤与众影响因素之间的作用机理,定量研究基于多种智能化数据挖掘方法的运营地铁结构健康风险感知与诊断具有重要意义。在综合各界学者研究理论的基础上,通过运营地铁结构损伤机理分析,提出基于局部病害、结构耐久性、渗透水、外部环境和其他五维度共涵盖20个内部指标的运营地铁结构健康风险评价指标体系;基于改进峰值云变换算法将指标属性划分为五个离散的区间,借助Rosetta软件导入双层决策表,基于遗传算法约简掉7个冗余属性,实现不确定条件下运营地铁结构健康知识发现;结合多维云理论和Copula函数理论,构建相依性条件下多维云-Copula模型,计算联合隶属度实现相依性条件下的结构健康风险感知;将联合隶属度分布转化为基本可信度分布,改进传统D-S证据理论进行两级多源信息融合,实现多传感器作用下结构健康风险评价,并结合蒙特卡罗模拟确定敏感性因素,为后期决策提供依据,增强措施针对性和管理效率。更多还原

【Abstract】 With the large-scale of urbanization and the expansion of urban rail transit network,the subway has become one of the most convenient way to travel.However,with the combined action of internal factors and external environment,kinds of problems appeared during the process of metro operation,which has attracted the attention of people and scholars.Large-scale construction and operation of subway started relatively later in China,for the lack of related research system and the ability to carry out structural health risk perception and diagnosis of operated subway on the basis of kinds of intelligent data mining methods is relatively weak.So it is of great significance to explore the mechanism between the damage and kinds of influencing factors of the subway,and study the structural health risk perception and diagnosis of operation subway on the basis of kinds of intelligent data mining methods.With the theory and study of kinds of scholars,the structure damage mechanism of operation subway has been analyzed,and an evaluation index system covering 20 internal indicators and five dimensions,which are the local disease,the structural durability,the permeable water,the external environment and the other,has been put forward.The continuous indicators are divided into five discrete intervals with the improved cloud model.Seven redundant indicators have been reduced with genetic algorithm on the basis of rough set and the software of Rosetta,to realize the structural health knowledge discovery of operation subway under uncertain conditions.The theory of multi-dimension cloud model and Copula function have been combined to establish the multidimensional cloud-Copula model and calculate joint membership degree to realize the structural health risk perception of operation subway under the condition of interdependence.At last,the joint membership degree has been transformed into the basic reliability distribution,and the traditional D-S evidence theory has been improved to carry out two-level multi-source information fusion,to get the structural health risk assessment of operation subway with the action of multi-sensor.Monte Carlo simulation has been used to get the sensitivity factors,which can provide basis for later decisions and enhance the effectiveness of management.更多还原